Fifty nine isolates of Ustilaginoidea virens, the cause of false smut of rice (Oryza sativa L.), were obtained from 46 rice hybrids in 14 counties in Sichuan, China during a survey conducted in 2006. Their pathogenicity was tested by inoculation on 3 rice hybrids, Gangyou182, Gangyou94- 11 and Yixiangyou2292, 6-9 days before heading in 2007. False smut was assessed 3 weeks after inoculation using a disease index (DI) based on symptom frequency on the panicles. The sporulation capacity of these isolates was also measured. The results showed that: (i) DIs were significantly different (P<0.01) both among pathogen isolates and rice hybrids, ranging from 0 to 98.52; (ii) there was a significant interaction between isolates and hybrids (P<0.01); (iii) significant differences in sporulation among the 59 isolates were found (P<0.01), but no relationship between sporulation and virulence on the 3 hybrids; (iv) variation in sporulation was observed among isolates originating from different counties and from the same county (P<0.05); (v) there were significant differences (P<0.01) between isolate groups from different host origin, female parents and male parents. Our results indicated a linkage between the pathogenicity of U. virens isolates and the resistance of rice hybrids. The 59 isolates could be classified into 6 groups based on their virulence to the tested rice hybrids. While variation in sporulation did not indicate host genotype-pathogen isolate interaction, pathogenicity data suggest specialisation, which is dependent on the site of origin of the isolates, the original host (rice hybrids) and the parentage of the original host.